How Machine Learning is Used in Store Assortment Optimisation

How Machine Learning is Used in Store Assortment Optimisation

December 04, 2022 | By HIVERY

In retail, assortment planning is the process of deciding which products to sell in a store, how much (space allocation), and what likely impacts the bottom line. It is also the process of deciding how to distribute the best new product innovation items and the impact of demand on other existing items (i.e., demand transfer). This process is usually done manually by a team of planners considering various factors such as sales data, store attributes, customer demographics, store location, and seasonality. However, with the advent of machine learning, it is now possible to automate the assortment planning process.

Introduction to Machine Learning 

Machine learning is a form of artificial intelligence that enables computer systems to learn from data, identify patterns, and make predictions without being explicitly programmed. It is used in many industries, including retail, to automate tasks, optimize processes, and make decisions. Machine learning algorithms are used to analyze large amounts of data to identify correlations and trends, which can be used to make better decisions and improve customer experiences.

What is Retail Assortment Planning?

Retail assortment planning is selecting and organizing products for a retail store. It involves analyzing customer preferences and buying habits through sales data to determine the best product assortment. Assortment planning helps retailers to maximize sales and profits by ensuring that the right products are available at the right time, in the right place, and at the right price.

How is Machine Learning Used in Retail Assortment Planning?

Machine learning is being used to improve the process of retail assortment planning. By leveraging large data sets and advanced algorithms, machine learning can help retailers to understand better customer preferences and buying patterns to create more effective assortments. Using machine learning, retailers can identify correlations between products, analyze customer behavior, and make better decisions about product selection. Additionally, machine learning can optimise pricing and promotions for each product, allowing retailers to maximise sales and profits, especially when launching new product innovations. For example, machine learning can inform retailers and suppliers which specific stores and how much to supply that product and the impact of demand transference to the category.

Machine learning algorithms can also improve forecasting, enabling retailers to predict customer demand better and ensure that the right products are available at the right time. By combining customer data with historical sales data, machine learning can provide insights into customer preferences and buying patterns, allowing retailers to make more accurate predictions about future demand.

Overall, machine learning is an invaluable tool for retailers looking to optimise their assortment planning process. By leveraging the power of machine learning, retailers can make more informed decisions that increase sales and profits.

The Challenges of Using Machine Learning in Retail Assortment Planning

Despite the potential benefits of machine learning in retail assortment planning, some challenges are still associated with its use. One of the main challenges is the lack of reliable data. Retailers need access to large amounts of data to train machine learning algorithms effectively. This can be difficult to obtain, as customer preferences and buying patterns vary greatly. Additionally, machine learning algorithms can be difficult to understand and interpret, making it difficult to make the right decisions based on the results.  This is why at HIVERY, we provide full transparency through the ability to view the “audit trail.” You can see how the machine learning model generated the results and how your business goals, merchandising rules, and constraints impacted the financial outcomes of your assortment. 

Another challenge is the cost associated with using machine learning. The hardware and software required to run machine learning algorithms can be expensive, and hiring experts to develop and maintain the algorithms is also expensive. Furthermore, there is the risk that the results of the algorithms may be incorrect or misinterpreted, leading to costly mistakes.  It is why at HIVERY, for instance, every engagement starts with “discovery and validation” after we provide the “first cut” of any results generated. This allows us to refine and fine-tune data sets, business goals, and merchandise rules.  We also generate what we call a “base planogram.”  Here we analyze and forecast clients' current planograms (i.e., PSA files), which the client validates, given they already know the results.

Finally, machine learning algorithms can be time-consuming to develop and maintain. Finding the right data sets and developing accurate algorithms can take time. Then, the ongoing work of monitoring and updating the algorithms as customer preferences and buying patterns change.


Machine learning can be used in retail assortment planning in several ways. For example, machine learning can predict product demand, identify trends, and optimise stock levels. Additionally, machine learning can be used to improve forecasting accuracy and reduce the need for manual intervention. When used correctly, machine learning can be a powerful tool for retailers looking to improve their assortment planning.

HIVERY offers an incredible assortment of strategy simulation and optimization technologies that help bolster decision-making. If you are looking for assortment optimization help, check out what we can do for you!

Related resources you might be interested in:

Blog: How To Get Assortment Space Contribution Index Optimized

Blog: Do Current Assortment Optimization Methods Create Store Sacrifices?

Blog: Demand Sensing: The New Era of Retail Demand Planning

Guide: Enabling Retail Space Planning With AI

Subscribe to HIVERY updates